Concurrent sexual partnerships-or partnerships which overlap in time-have been suggested as a significant contributing factor in STI/HIV transmission and acquisition. However, most studies of concurrency have utilized cross-sectional study designs with retrospective recall of behavior. As a result, little is known regarding factors associated with the initiation of concurrent partnerships, and very few studies have assessed the quality of concurrent sexual partnership data. Different motivations for initiating concurrent sexual partnerships have been associated with different degrees of STI/HIV risk. Retrospective recall is subject to known methodological challenges including memory failure and imprecision; as concurrency is typically defined by an overlap in self- reported dates of partnership, establishing the degree of reporting error in sexual partnership dates has important implications for the classification of concurrency in both research and public health practice. We will analyze data from a racially and ethnically diverse sample of young adults aged 18-30 and their sexual partners studied in the Project on Partner Dynamics (POPD) to examine to what extent: 1) a participant's belief that his/her partner has another sexual partner (perceived partner concurrency) predicts the participant's initiation of a concurrent partnership during the following 4 months; 2) sexual partners report the same dates for their first sexual intercourse (anal, oral, or vaginal) and for their last sexual intercourse; and 3) the observed association between perceived partner concurrency and subsequent concurrency initiation is affected by potential error in recall of dates. The POPD study recruited 537 persons from Los Angeles area community sites for a one-year study of heterosexual partnering, including interviewer-administered questionnaires about sexual risk perceptions, intentions, and behaviors at baseline, 4 months, 8 months, and 12 months. At 8- and 12-months, participants were invited to bring a sexual partner for interview. Index and partner respondents were interviewed separately and also jointly, to obtain couple-level answers on selected survey iteMs. We will use generalized estimating equations (GEE) to estimate the association between perceived partner concurrency and subsequent initiation of concurrent behavior over the following 4-months. We will compare reported dates of first and last sexual encounters within the partnership by examining partnership-level agreement, including percent agreement, weighted kappa, and the distribution of differences in reported dates. We will use the observed distribution of differences in a sensitivity analysis to investigate the impact of potential reporting error on the classification of concurrency and on the observed association between perception of partner concurrency and subsequent initiation of concurrency. This information on the quality of concurrency measurement has important implications for STI research. Knowledge about predictors of concurrency initiation may provide insights into addressing STI disparities among racial/ethnic minorities.